Novel Predictive Fuzzy Logic-Based Energy Management System for Grid-connected and Off-grid Operation of Residential Smart Micro-grids

IEEE In this paper, a novel energy management system with two operating horizons is proposed for a residential micro-grid application. The micro-grid utilises the energies of a photovoltaic (PV), a fuel cell and a battery bank to supply the local loads through a combination of electric and magnetic buses. The proposed micro-grid operates in a large number of grid-connected and off-grid operation modes. The energy management system includes a long-term data prediction unit based on a 2D dynamic programming and a short-term fuzzy controller. The long-term prediction unit is designed to determine the appropriate variation range of the battery state of charge and fuel cell state of hydrogen. The efficiency performance of the micro-grid components, predicted energy generation and demand, energy cost and the system constraints are taken into account. The resultant data then is sent to the short-term fuzzy controller which determines the operation mode of the micro-grid based on the real-time condition of the micro-grid elements. A prototype of the proposed micro-grid including the energy management system is developed, and experimental tests are conducted for three different energy management scenarios. The proposed management technique is validated through energy distribution and cost analysis